Vol. 63, No. 5
Discrimination of Virulent and Avirulent Streptococcus suis
Capsular Type 2 Isolates from Different Geographical Origins
SYLVAIN QUESSY,* J. DANIEL DUBREUIL, MARTINE CAYA,
Received 1 October 1993/Returned for modiﬁcation 16 November 1993/Accepted 24 January 1994
cause of septicemia and meningitis in swine (1). It can also
induce clinical manifestations in humans (2). Little is known
about the pathogenesis of the infection. Williams (16) has
reported that virulent strains could survive within macro-
phages. Vecht et al. (15) have described for European isolates
a 110-kDa extracellular factor (EF) and a 136-kDa cell wall
protein previously known as the muraminidase-released pro-
tein (MRP). These proteins were present only in strains viru-
lent for pigs and therefore were reported to be virulence mark-
ers. On the basis of the presence of MRP and EF in the culture
supernatants of the strains, three phenotypes were described:
, virulent strains; MRP
with slight pathological changes; and MRP
strains (15). The same authors subsequently developed a dou-
ble-antibody sandwich enzyme-linked immunosorbent assay
with monoclonal antibodies directed against those two viru-
lence markers to discriminate virulent and avirulent isolates
the epidemiology of the S. suis capsular type 2 infections (1).
Since vaccination often leads to equivocal results, the detection
of animals carrying virulent strains could be very helpful in the
prevention of the infection (1). Furthermore, the characteriza-
tion of virulent and avirulent isolates, the identiﬁcation of
virulence determinants, and the development of an experimen-
tal model of infection are important steps towards understand-
ing of the pathogenesis of the infection (10). Mouse models
have proven to be an important tool in studying S. suis capsular
type 2 infections, allowing the evaluation of bacterial virulence
(3, 8, 17).
In this study, we have compared, using Western blots (im-
munoblots), the immunogenicities of proteins from the cellular
fractions and the culture supernatants of S. suis serotype 2
strains from different geographical origins with the aim of
relating the protein proﬁle to the virulence of the strains. We
identiﬁed a virulence marker and evaluated its protective effect
by using the experimental mouse model of infection.
MATERIALS AND METHODS
Bacterial strains and growth conditions.
Twenty S. suis capsular type 2 isolates
were used in this study. The serotype 2 reference strain (735), isolated in Den-
mark, was provided by J. Henrichsen, Statens Seruminstitut, Copenhagen, Den-
* Corresponding author. Phone: (514) 773-7730, ext. 101. Fax: (514)
TABLE 1. Evaluation of the virulence of S. suis capsular type 2
strains from different origins in an experimental murine model
HV, highly virulent (seven or more deaths); MV, moderately virulent (three
to six deaths); AV, avirulent (fewer than three deaths) (3).
T. J. L. Alexander, Department of Clinical Veterinary Medicine, University of
Cambridge, Cambridge, United Kingdom. Isolates from the United States,
AAH4 and 614, were provided by Brad Fenwick, Kansas State University, Man-
hattan. Mexican isolates, JL590 and JL819, were provided by Jose Luis Monter
Flores, University of Toluca, Toluca, Mexico. Isolates from The Netherlands,
AR770297 and AR770357, were provided by J. P. Arends, Groningen, The
Netherlands. The other 11 isolates were from the Faculty of Veterinary Medi-
cine, University of Montreal, St-Hyacinthe, Canada. With the exceptions of
strains 4/40 H2, 4/3 H1, and 4/39 H1 (from healthy pigs), AR770297 and
AR770357 (from human meningitis cases), and 741 (from bovine abortion), all
strains originated from diseased pigs.
For each strain, four colonies from a 24-h culture on 5% bovine blood agar
plates were inoculated in 200 ml of Todd-Hewitt broth and incubated overnight
in an aerobic atmosphere at 37
ЊC. Cells were harvested by centrifugation, washed
with a sterile saline solution, and resuspended in 3 ml of K
(0.1 M, pH
7.0). Culture supernatants were collected and concentrated 100 times by ultra-
ﬁltration (type YM 30 ﬁlters; Amicon Corp., Danvers, Mass.).
The virulence of the non-
Canadian isolates (except strains 735 and AR770297) was estimated by using a
mouse model already described (3). The virulence of the Canadian isolates
(except strain 0891) had already been evaluated with that model (3). Brieﬂy,
strains were growth in Todd-Hewitt broth supplemented with inactivated bovine
serum to an optical density of 0.1 (at 540 nm), 1 ml of the suspension was injected
intraperitonealy into groups of 10 28-day-old CF1 mice, and mortality was re-
corded for 1 week. In order to detect any toxic effect of the culture supernatant,
groups of 10 mice were also injected with 100-fold-concentrated supernatants of
overnight cultures of strains 1591 and 735. To demonstrate the relevance of the
model of infection in the natural host, four virulent isolates (1591, 999, AAH4,
and JL590) and four avirulent isolates (1330, TD10, R75/S2, and 0891) were
injected into groups of three 6-week-old cross-bred pigs by the same protocol but
by the intravenous route. Pigs had previously been tested by using the Western
blot technique to detect antibodies against S. suis and were monitored twice a
day for 10 days following the experimental infection.
Production of antisera.
Adult New Zealand rabbits were injected once a week
for 4 weeks intramuscularly with a formalin-killed culture (0.5% [vol/vol] forma-
lin was added to an overnight culture in Todd-Hewitt broth) of the reference
strain. In order to obtain antisera from speciﬁc fractions of the cellular protein
proﬁle, other rabbits were injected once a week with polyacrylamide gel bands of
128 and 135 kDa from the cellular protein proﬁle (see below) mixed with
incomplete Freund’s adjuvant. Bands corresponding to the 128-kDa cell fraction
were collected from the processed strain 1591 culture, while the cellular fraction
bands of 135 kDa were collected from the reference strain. Monoclonal anti-
bodies raised against the 136-kDa MRP and the 110-kDa extracellular protein
(15) were kindly provided by U. Vecht (DLO-Central Veterinary Institute, Lely-
stad, The Netherlands).
Cells from the different S. suis strains were processed in a
French press cell (20,000 lb/in
, three times), treated with lysozyme (Sigma
ЊC, and centrifuged (12,500
ϫ g, 20 min). Supernatants were harvested, and the protein content was evalu-
ated by a Bradford colorimetric assay (Bio-Rad, Hercules, Calif.). This solution,
as well as the concentrated culture supernatants, was mixed with an equal volume
of solubilization buffer, boiled for 4 min, and processed in 5 and 7.5% polyacryl-
amide vertical slab gels (with a 4.5% stacking gel) as described previously (9).
Following sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-
PAGE), material was transferred from the slab gel to the nitrocellulose mem-
brane by the methanol-Tris-glycine system (12). Electroblotting was done in a
Transblot apparatus (Hoefer Scientiﬁc Instruments, San Francisco, Calif.) for 18
h at 60 V. Casein (2%, wt/vol) was then used to block unreacted sites, and the
nitrocellulose membrane was incubated for 2 h with 1:200 (vol/vol) dilutions of
FIG. 1. Western blots of different S. suis capsular type 2 strains. SDS-PAGE (5.0%) was performed with culture supernatants. Protein proﬁles were revealed with
rabbit antiserum raised against the reference strain of serotype 2. Left, molecular mass markers (in kilodaltons). Bottom, strain identiﬁcation numbers. Right, large
arrowhead, 135-kDa protein; small arrowhead, 110-kDa protein.
TABLE 2. Experimental infection of pigs with S. suis serotype
A, F, LD, N (Eu
A, F, N
A, F, LD (Eu)
A, F, LA, VD (Eu)
A, F, N
A, F, LA
A, F, VD (Eu)
See Table 1, footnote b.
Three pigs were injected intraveneously with 3
CFU of each S. suis
experiments, which had similar results.
As recorded during the 10-day experiment. A, anorexia; F, fever; LA, lame-
ness; LD, persistent lateral decubitus; N, nervous signs; VD, persistant ventral
decubitus; —, no signs.
Main pathological lesions.
Eu, animal showed persistant decubitus for more than 12 h and was eutha-
with a peroxidase-labeled immunoglobulin G (IgG) fraction of goat antiserum
raised against rabbit IgGs (Bio-Rad) for 60 min at a dilution of 1:5,000 in a 2%
casein in Tris-NaCl. After repeated washings, the presence of bound antigens
was visualized by reacting the nitrocellulose membrane with 0.06% 4-chloro-1-
naphthol (Sigma) in cold methanol mixed with 0.02% H
parent molecular masses were calculated by comparison with standards of known
Immunization assays with the 110-kDa fraction.
The concentrated culture
supernatant of strain 1591 was processed on a polyacrylamide gel and stained
with Coomassie blue. The 110-kDa band was excised from the gel, mixed with
Freund’s incomplete adjuvant as previously described (6), and injected into a
group of 13 mice once a week for 3 weeks, with each mouse receiving approxi-
g of protein at each injection. On day 21, three mice were euthanized,
and their blood was collected. The IgG response to the protein was monitored by
Western blot, and this speciﬁc antiserum was used to detect the 110-kDa protein
in the different strains. The other 10 mice were experimentally infected with the
reference strain 735 (5 mice) and strain 1591 (5 mice).
By using the mouse experimental model of infection, S. suis
strains were classiﬁed as highly virulent, moderately virulent,
or avirulent (Table 1). The three most virulent isolates were
strains 1591 and 999 (Canadian isolates) and JL590 (a Mexican
isolate), while four other strains were found to be avirulent
(strains 0891 and 1330 [Canada] and TD10 and R75/S2 [Unit-
ed Kingdom]). When the four avirulent strains and four viru-
lent strains, as estimated with the mouse model of infection,
were injected into pigs, all virulent strains reproduced the
disease and all avirulent strains failed to reproduce the disease
(with the exception of a transient lameness in one case and
anorexia and fever in another one) (Table 2). S. suis was
recovered from lesions of all but one of the diseased pigs. The
FIG. 2. Western blots of different S. suis capsular type 2 strains. SDS-PAGE (5.0%) was performed with culture supernatants. Protein proﬁles were revealed with
rabbit antiserum raised against the 135-kDa protein. Left, molecular mass markers (in kilodaltons). Bottom, strain identiﬁcation numbers. Arrowhead, 135-kDa protein.
FIG. 3. Western blots of different S. suis capsular type 2 strains. SDS-PAGE (7.5%) was performed with cellular fractions. Protein proﬁles were revealed with rabbit
antiserum raised against the reference strain. Left, molecular mass markers (in kilodaltons). Bottom, strain identiﬁcation numbers. Right, large arrowhead, 135-kDa
protein; small arrowhead, 128-kDa protein.
VIRULENT AND AVIRULENT S. SUIS ISOLATES
735 failed to induce any clinical signs in mice.
When the protein proﬁles of the culture supernatants were
compared, a protein of about 110 kDa was found to be present
in all moderately and highly virulent strains (except the human
isolates AR770357 and AR770297, from The Netherlands) and
to be absent in all avirulent isolates (Fig. 1). Variation in the
molecular mass of this protein was noted. Western blotting
performed with antisera of mice immunized against the 110-
kDa supernatant protein from strain 1591 recognized the dif-
ferent variants. A protein of about 135 kDa was detected in all
virulent and avirulent strains except strains 1591 and 999 (Fig.
1) and 6891. Furthermore, the rabbit antiserum produced
against the 135-kDa band did not detect any protein in the
culture supernatants of strains 1591 and 999 (Fig. 2) and 6891,
while it did in other strains.
When the cellular protein proﬁles of the different isolates
were compared by using Western blots and rabbit antisera, the
protein proﬁles of the highly and moderately virulent strains of
different origins were similar, with the exception of three Ca-
nadian isolates, strains 1591 and 999 (Fig. 3) and 6891, in which
a 135-kDa protein was not detected. The 135-kDa protein was
present in all avirulent strains tested. Western blots of the cell
protein fraction with antiserum produced against the 135-kDa
protein did not detect this protein in the cellular fractions of
strains 1591 and 999 (Fig. 4A) and 6891 but detected it in all
the other strains tested. When antiserum raised against the
128-kDa fraction was used, Western blots showed that this
antiserum could recognize the 128-kDa protein and to a lesser
extent the 135-kDa protein in all strains except strains 1591
and 999 (Fig. 4B) and 6891, in which only the 128-kDa protein
With monoclonal antibodies raised against the 136-kDa
MRP, results were similar to those with polyclonal antibodies
raised against the 135-kDa protein. We did not detect the
136-kDa protein in the supernatants of strains 1591, 999, and
6891 even when the supernatant was 100-fold concentrated.
The monoclonal antibodies raised against the 110-kDa EF did
not recognize the 110-kDa protein of the strains used in this
from the culture supernatant of strain 1591 showed that this
protein induced an IgG response and protected mice against
infection with the homologous strain and even with another
virulent strain (735) (Table 3).
FIG. 4. Western blots of different S. suis capsular type 2 strains. SDS-PAGE (7.5%) was performed with cellular fractions. Protein proﬁles were revealed with rabbit
antiserum raised against the 135-kDa protein (A) and the 128-kDa protein (B). Left, molecular mass markers (in kilodaltons). Bottom, strain identiﬁcation numbers.
Right, large arrowheads, 135-kDa protein; small arrowhead, 128-kDa protein.
TABLE 3. Protective effect of a 110-kDa protein of S. suis capsular
type 2 against experimental infection in mice
Presence of 110-kDa
band on Western
blot (no. positive
Mice were injected with the 110-kDa protein or phosphate-buffered saline
(PBS), each in Freund’s incomplete adjuvant.
Means of two separate experiments with similar results.
QUESSY ET AL.
The identiﬁcation and characterization of virulence deter-
minants can be very important for the understanding of the
pathogenesis of an infection. Vecht et al. (15) reported that
both a membrane protein of 136 kDa (MRP) with homology to
a Staphylococcus aureus ﬁbronectin-binding protein (11) and a
110-kDa EF were virulence markers for S. suis capsular type 2
isolates. In this study, a 135-kDa protein, or a variant of about
this molecular mass, has been found to be present in the
majority of virulent isolates but also in avirulent isolates. On
the other hand, an EF of about 110 kDa has been found to be
present in all virulent strains, with the exception of two human
isolates, and absent in the avirulent isolates. However, the
110-kDa protein found in this study was antigenically different
from the EF previously reported (15), as shown by the use of
monoclonal antibodies. Considerable genetic diversity has
been found among isolates of the same S. suis serotype (5), and
it is thus possible that phenotypic variants of the 110-kDa
protein might exist. We are currently investigating the related-
ness of these two proteins. Variability in the molecular mass of
the 110-kDa protein was found in this study. The use of speciﬁc
antisera showed that these proteins were related. The fact that
strains 1591 and AAH4, with 110-kDa proteins showing vari-
ation in molecular mass, both produced disease in pigs tends to
show that both forms of the 110-kDa protein found as a viru-
lence marker in this study can be recovered from virulent
strains. The possible role of this protein is not yet known, but
the fact that mice injected with concentrated supernatant did
not show any clinical signs would indicate that it is not by direct
activity that this protein would play a role in the infection.
By using a rabbit antiserum raised against the reference
strain, the 135-kDa protein was not detected in two highly
virulent Canadian strains. However, in those two strains, a
128-kDa protein was present, and the use of antiserum pro-
duced against that protein revealed the presence of cross-
reactivity with the 135-kDa band. This suggests that these pro-
teins could be antigenically related to one another. However,
the 135-kDa fraction was not detected in the cellular fractions
or in the culture supernatants of these two highly virulent
strains by using a 135-kDa-speciﬁc antiserum or an antiserum
raised against the whole bacterial cell. The existence of highly
virulent strains that do not possess the 135-kDa protein raised
some questions about the role of this protein in S. suis infec-
mental model of infection (3, 8, 17) as well as in the natural
host (13) have been noted. The TD10 and R75/S2 isolates were
previously shown to possess a low level of virulence (6) and
were found to be avirulent in mice in the present study. In
reproducing the disease in the natural host, some authors (15)
had to use preinfection with Bordetella spp. even for strains
reported to be virulent. Experiments carried out with pigs in
the present study have conﬁrmed the presence of differences in
the virulence of strains; they have also demonstrated the value
of the mouse experimental model used (3). Finally, in accor-
dance with the results of Iglesias et al. (7), it has been shown
that highly virulent isolates could experimentally produce the
disease in the natural host without the need for a preinfection.
Successful passive immunizations in mice with rabbit anti-
sera directed against cell wall protein have been reported (4,
6). The active immunization experiment reported here with a
110-kDa culture supernatant protein in mice showed that this
protein could induce an IgG response and adequately protect
against experimental infection with virulent homologous and
heterologous S. suis strains. Other experiments with the natu-
ral host are needed to conﬁrm the protective potential of this
fraction, but the 110-kDa protein might eventually be a good
candidate for a subunit vaccine.
We thank Rene
´tourneau for her excellent technical assistance.
pour la Formation de Chercheurs et l’aide a
` la Recherche. This work
was supported in part by grants from the Conseil des Recherches en
ˆche et en Agro-Alimentaire du Que
from the Conseil de Recherches en Sciences Naturelles et en Ge
Canada to R.H. (no. 2638).
1. Alexander, T. J. L. 1992. Streptococcus suis: an update. Pig Vet. Soc. Proc.
in humans. Rev. Infect. Dis. 10:131–137.
3. Beaudoin, M., R. Higgins, J. Harel, and M. Gottschalk. 1992. Studies on a
murine model for evaluation of virulence of Streptococcus suis capsular type
2 isolates. FEMS Microbiol. Lett. 99:111–116.
4. Gottschalk, M., R. Higgins, M. Jacques, and D. Dubreuil. 1992. Production
and characterization of two Streptococcus suis capsular type 2 mutants. Vet.
5. Hampson, D. J., D. J. Trott, L. Clarke, C. G. Mwaniki, and I. D. Robertson.
1993. Population structure of Australian isolates of Streptococcus suis. J.
Clin. Microbiol. 31:2895–2900.
6. Holt, M. E., M. R. Enright, and T. J. L. Alexander. 1990. Protective effects
of sera raised against different fractions of Streptococcus suis type 2. J. Comp.
7. Iglesias, J. G., M. Trujano, and J. Xu. 1992. Inoculation of pigs with Strep-
tococcus suis type 2 alone or in combination with pseudorabies virus. Am. J.
Vet. Res. 3:364–367.
8. Kataoka, Y., M. Haritani, M. Mori, M. Kishima, C. Sugimoto, M. Nakazawa,
and K. Yamamoto.
1991. Experimental infections of mice and pigs with
9. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of
the head of bacteriophage T4. Nature (London) 227:680–685.
10. Smith, H. 1988. The development of studies on the determinants of bacterial
pathogenicity. J. Comp. Pathog. 98:253–273.
11. Smith, H. E., U. Vecht, A. L. J. Gielkens, and M. A. Smits. 1992. Cloning and
nucleotide sequence of the gene encoding the 136-kilodalton surface protein
(muramidase-released protein) of Streptococcus suis type 2. Infect. Immun.
12. Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoretic transfer of
proteins from polyacrylamide gels to nitrocellulose sheets: procedure and
some applications. Proc. Natl. Acad. Sci. USA 76:4350–4354.
13. Vecht, U., J. P. Arends, E. J. van der Molen, and L. A. G. van Leengoed. 1989.
Differences in virulence between two strains of Streptococcus suis type II
after experimentally induced infection of new born germ-free pigs. Am. J.
Vet. Res. 50:1037–1043.
14. Vecht, U., H. J. Wisselink, J. Anakotta, and H. E. Smith. 1993. Discrimina-
tion between virulent and non-virulent Streptococcus suis type 2 strains by
enzyme-linked immunosorbent assay. Vet. Microbiol. 34:71–82.
15. Vecht, U., H. J. Wisselink, M. L. Jellema, and H. E. Smith. 1991. Identiﬁ-
cation of two proteins associated with virulence of Streptococcus suis type 2.
Infect. Immun. 59:3156–3162.
16. Williams, A. E. 1990. Relationship between intracellular survival in macro-
phages and pathogenicity of Streptococcus suis type 2 isolates. Microb.
17. Williams, A. E., W. F. Blakemore, and T. J. L. Alexander. 1988. A murine
model of Streptococcus suis type 2 meningitis in the pig. Res. Vet. Sci.
. 63, 1995